6  Merge ETL

Using the ETL generated based on this document, one target database per hospital will be created. To allow studies done on a combined data set, a ‘merge ETL’ would have to be written. The aim of this ETL is to combine the outputs of the different datasets into one big target database. For this, people would need to be deduplicated, as they could have person and death records from multiple hospitals which is in violation with OMOP conventions. A person and death entry should be unique. Once the person and death tables have been deduplicated and the person_ids were updated, this needs to be reflected in the other (clinical) tables. Lastly, when adding all events together there will be an overlap in the ids, these will have to be updated to ensure unique values are being used.